Messages in this thread | | | Date | Tue, 14 Apr 1998 20:01:10 -0400 (EDT) | From | "Benjamin C.R. LaHaise" <> | Subject | RE: Improving Ext2 Reliability |
| |
On Tue, 14 Apr 1998, Jordan Mendelson wrote:
> The cache takes quite a long time to rebuild and I don't feel like wasting > bandwidth to save 30 minutes :)
You've missed a major concern: a cache with corrupt data is a disservice to your users. Squid does *not* verify that the data contained within the cache is correct. Some very good points to this effect have been made in the recent thread 'NT causes $10M loss' on cold.system. Consistent metadata buys you nothing if you don't verify the data's valid. Fsck is *only* meant to exit when a filesystem has valid metadata which the sysadmin can use to clean up the mess after a crash (preferably restoring from backup).
> I'm finally just scrapping 2.0.x and moving to 2.1.9x so I can at least get > alt-sysrq and some better debugging messages. Hell, 2.1.9x has been more > stable on my development boxes than the 2.0.33+ on my production machines. > :)
Good luck! =)
-ben
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu
| |